Siemens Teams Up in OT Endpoint Security

Industrial control system (ICS)/SCADA equipment giant Siemens today made its first foray into the endpoint security space with a new collaboration product and service offering aimed at devices sitting in industrial operational technology (OT) networks. 

Siemens and machine learning (ML) endpoint security vendor SparkCognition teamed up on an ML-based endpoint protection product and service for devices on plant networks. The so-called DeepArmour Industrial software initially runs on Windows-based devices, and it was built off of SparkCognition's enterprise version of the endpoint security tool for detecting known and previously unknown threats. Siemens will provide custom integration, installation, and security alert analysis for plant networks.

"We'll take alarms into our environment, analyze them, and help them think about how to respond to any threats," says Leo Simonovich, head of industrial cybersecurity at Siemens. "The way we think about this is as a technology-delivery-as-a-service," he says.

Siemens started doubling down on managed security services three years ago, adding third-party security vendor technologies to its networking monitoring and security services such as DarkTrace, PAS, and Tenable. Security services, especially cloud-based ones, provide new sources of revenue for legacy ICS vendors such as Siemens.

Simonovich says Siemens' endpoint partnership with SparkCognition aims to address a longstanding problem of tracking and protecting endpoints in OT networks. The goal is to give OT networks more visibility into endpoint devices running on them and any threats that land on or originate from those devices.

The product initially supports only endpoints running Windows 2000 and later versions, which includes programmable logic controller HMIs (human machine interfaces) on those platforms. The plan is to ultimately add support for non-Windows and Windows-embedded platforms (including HMIs), both Siemens and non-Siemens equipment.

Patching software in an OT network is a dilemma, from the perspectives of both disruption of operations and software so old that it has no updates. "Many of these [OT] assets are not maintained" and patched for decades, Simonovich says, so they're left exposed to attack, especially if they are located in remote physical locations.

Ernie Hayden, an ICS consultant with 443 Consulting Ltd., says most HMIs he sees in factories and plants run on older Windows platforms such as XP and Windows 2000. "It would be interesting it that can back-fit into those operating systems," he says of the Siemens-SparkCognition offering.

SparkCognition CTO Sridhar Sudarsan says his company is building ML models for the endpoint product to detect and thwart attacks. "We're using ML models to train and build features — multiple models based on millions of zero-day data collected over time."

OT networks have increasingly been facing malware infections carried into the plant via endpoints. A Ponemon Institute study commissioned by Siemens in 2017 found that 68% of global energy firms had been hit with at least one cyberattack, and 67% say the risk to their systems has increased significantly.  

Researchers have tested the waters to see just what the bad guys are after, using honeypots. For seven months last year, researchers at Trend Micro ran a legitimate-looking phony industrial prototyping company with an advanced interactive honeypot network to attract would-be attackers. This honeypot-on-steroids successfully lured the same types of threats that IT networks face: ransomware, remote access Trojans, malicious cryptojacking, and botnet-style beaconing malware that infected its robotics workstation for possible lateral movement.

Most recently, Cybereason ran an ICS honeypot that got hit with multiple ransomware attacks, stealing data and user credentials as well.

Related Content

Learn from industry experts in a setting that is conducive to interaction and conversation about how to prepare for that "really bad day" in cybersecurity. Click for more information and to register.